WO2020124988A1 - Vision-based parking space detection method and device - Google Patents

Abstract

A vision-based parking space detection method and device. The method comprises: preprocessing a captured environmental image on two sides of a vehicle to obtain a corresponding bird's-eye view (101); detecting a parking space frame and a parking space corner point in the bird's-eye view (102) and selecting a target parking space frame from the parking space frame in the bird's-eye view (103); acquiring, according to the target parking space frame, the target parking space corner point on the target parking space frame from the parking space corner point (104); determining, according to the position coordinates of the target parking space corner point in the bird's eye view, the real position coordinates corresponding to the target parking space corner point (105); and determining, according to the real position coordinates corresponding to the target parking space corner point, the target parking space corresponding to the target parking space frame (106). By means of this technical solution, the accuracy of parking space detection can be improved, thereby improving the recognition rate of parking spaces.

Description

Vision-based parking space detection method and device Technical field

The invention relates to the technical field of parking space detection, in particular to a vision-based parking space detection method and device.

Background technique

In recent years, with the rapid economic development, the number of cars has increased rapidly, but the number of parking spaces is limited, which brings some difficulties to find parking spaces. The development of parking space detection technology has solved the problem of finding parking spaces to a certain extent. problem. At present, there are several methods of parking space detection: the method of identifying parking spaces based on color segmentation technology, the method of identifying parking spaces based on Canny edge detection and Hough line detection technology, and the method of identifying parking spaces based on SIFT feature extraction and matching technology. However, the above-mentioned method of identifying the parking space based on the color or shape of the parking space line has a low degree of recognition, and the success rate of identifying the parking space is not high. Therefore, there is an urgent need to develop a parking space detection method with a high parking space recognition rate.

Summary of the invention

The embodiment of the invention discloses a vision-based parking space detection method and device, which can improve the accuracy of the parking space detection, thereby improving the recognition rate of the parking space.

A first aspect of an embodiment of the present invention discloses a vision-based parking space detection method, including:

Preprocess the environmental images on both sides of the captured vehicle to obtain the corresponding bird's-eye view;

Detecting a parking space frame and a parking space corner point in the bird's-eye view;

Selecting a target parking space frame from the parking space frame in the bird's-eye view;

Obtaining the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame;

Determine the real position coordinates corresponding to the target parking space corner point according to the position coordinates of the target parking space corner point in the bird's-eye view;

The target parking space corresponding to the target parking space frame is determined according to the real position coordinates corresponding to the target parking space corner point.

As an optional implementation manner, in the first aspect of the embodiments of the present invention, the preprocessing the captured environmental images on both sides of the vehicle to obtain a corresponding bird's-eye view includes:

According to the internal parameters and external parameters of the camera's camera device, perform distortion correction on the captured environmental images on both sides of the vehicle to obtain corresponding corrected images;

Perform perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view.

As an optional implementation manner, in the first aspect of the embodiments of the present invention, distortions are corrected for the captured environmental images on both sides of the vehicle according to the internal and external parameters of the camera's camera device, Before obtaining the corresponding corrected image, the method further includes:

Obtain the internal and external parameters of the camera's camera device; the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the camera Device rotation matrix;

According to the internal parameters and external parameters of the camera's camera device, the captured environmental images on both sides of the vehicle are subjected to distortion correction to obtain corresponding corrected images, including:

According to the internal parameter and the external parameter, determine a first correspondence between the environment images on both sides of the vehicle and corresponding corrected images; the first correspondence includes the captured environment on both sides of the vehicle Correspondence between pixels of the image and pixels of the corrected image;

Save the first correspondence to the first lookup table;

According to the first look-up table, perform distortion correction on the captured environmental images on both sides of the vehicle to obtain the corrected image;

The performing perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view includes:

According to the pixels of the corrected image, determine a second correspondence between the corrected image and the corresponding bird's-eye view; the second correspondence includes the pixels of the corrected image and the pixels of the bird's-eye view Correspondence between

Save the second correspondence to the second lookup table;

According to the second lookup table, perform a perspective transformation on the corrected image to obtain the bird's-eye view.

As an optional implementation manner, in the first aspect of the embodiments of the present invention, before the preprocessing the captured environmental images on both sides of the vehicle to obtain a corresponding bird's-eye view, the method further includes:

Obtain the internal and external parameters of the camera's camera device; the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the camera Device rotation matrix;

And, the preprocessing of the environmental images on both sides of the captured vehicle to obtain a corresponding bird's-eye view includes:

According to the internal parameter and the external parameter, determine a third correspondence between the environment images on both sides of the vehicle and the corresponding bird's-eye view; the third correspondence includes the captured environment on both sides of the vehicle Correspondence between pixels of the image and pixels of the bird's-eye view;

Save the third correspondence to the third lookup table;

According to the third lookup table, perform a bird's eye view transformation on the captured environment images on both sides of the vehicle to obtain the bird's eye view.

As an optional implementation manner, in the first aspect of the embodiment of the present invention, after obtaining the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame, And before determining the real position coordinates corresponding to the target parking space corner point according to the position coordinates of the target parking space corner point in the bird's-eye view, the method further includes:

Detecting whether there is an obstacle in the area framed by the target parking space frame;

If there is an obstacle in the area framed by the target parking space frame, mark the area framed by the target parking space frame as a non-parking space;

If there is no obstacle in the area framed by the target parking space frame, the step of determining the real position coordinates corresponding to the target parking space corner point according to the location coordinates of the target parking space corner point in the bird's-eye view is performed .

A second aspect of an embodiment of the present invention discloses a parking space detection device, including:

The pre-processing unit is used to pre-process the environment images on both sides of the vehicle to obtain the corresponding bird's-eye view;

A first detection unit, configured to detect a parking space frame and a parking space corner point in the bird's-eye view;

A screening unit for screening a target parking space frame from the parking space frame in the bird's-eye view;

An obtaining unit, configured to obtain a target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame;

A first determining unit, configured to determine the real position coordinates corresponding to the target parking space corner point according to the position coordinates of the target parking space corner point in the bird's-eye view;

The second determining unit is configured to determine the target parking space corresponding to the target parking space frame according to the real position coordinates corresponding to the target parking space corner point.

As an optional implementation manner, in the second aspect of the embodiments of the present invention, the preprocessing unit includes:

The distortion correction subunit is used to perform distortion correction on the captured environmental images on both sides of the vehicle according to the internal parameters and external parameters of the camera device of the vehicle to obtain corresponding corrected images;

The perspective transformation subunit is configured to perform perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view.

As an optional implementation manner, in the second aspect of the embodiments of the present invention, the parking space detection device further includes:

The first parameter acquisition unit is used to perform distortion correction on the captured environmental images on both sides of the vehicle according to the internal and external parameters of the camera's camera device of the distortion correction sub-unit to obtain the corresponding corrected image To obtain the internal and external parameters of the camera's camera device; the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the The rotation matrix of the camera device;

The distortion correction subunit includes:

A first determining module, configured to determine the first correspondence between the environmental images on both sides of the vehicle and the corresponding corrected images based on the internal parameters and the external parameters; the first correspondence includes the captured Correspondence between pixels of the environmental images on both sides of the vehicle and pixels of the corrected image;

A first saving module, configured to save the first correspondence to the first lookup table;

A distortion correction module, configured to perform distortion correction on the captured environmental images on both sides of the vehicle according to the first lookup table to obtain the corrected image;

The perspective transformation subunit includes:

A second determining module, configured to determine a second correspondence between the corrected image and the corresponding bird's-eye view according to the pixels of the corrected image; the second correspondence includes the pixels of the corrected image and all The correspondence between the pixels of the bird's-eye view;

A second saving module, configured to store the second correspondence relationship in a second lookup table;

The perspective transformation module is configured to perform perspective transformation on the corrected image according to the second lookup table to obtain the bird's-eye view.

As an optional implementation manner, in the second aspect of the embodiments of the present invention, the parking space detection device further includes:

A second parameter acquisition unit, configured to acquire the internal parameters and external parameters of the camera device of the vehicle before the preprocessing unit preprocesses the captured environment images on both sides of the vehicle to obtain the corresponding bird's-eye view; The internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the rotation matrix of the camera device;

The pre-processing unit includes:

A determination subunit, configured to determine a third correspondence between the environmental images on both sides of the vehicle and the corresponding bird's-eye view based on the internal parameters and the external parameters; the third correspondence includes the A correspondence between pixels of environmental images on both sides of the vehicle and pixels of the bird's-eye view;

A saving subunit, configured to save the third correspondence to a third lookup table;

The bird's eye view transformation subunit is configured to perform bird's eye view transformation on the captured environment images on both sides of the vehicle according to the third lookup table to obtain the bird's eye view.

As an optional implementation manner, in the second aspect of the embodiments of the present invention, the parking space detection device further includes:

The second detection unit is used for the acquisition unit to acquire the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame, and the first determination unit according to the Detecting the position coordinates of the target parking space corner point in the bird's-eye view, before determining the real position coordinates corresponding to the target parking space corner point, detecting whether there is an obstacle in the area within the target parking space frame;

A marking unit, configured to mark the area in the target parking space frame as a non-parking space when the second detection unit detects that there is an obstacle in the area in the target parking space frame;

The first determining unit is specifically configured to determine the position of the corner point of the target parking space in the bird's-eye view when the second detection unit detects that there is no obstacle in the area framed by the target parking space frame Coordinates, determine the real position coordinates corresponding to the target parking space corner point.

A third aspect of an embodiment of the present invention discloses a parking space detection device, including:

Memory storing executable program code;

A processor coupled to the memory;

The processor calls the executable program code stored in the memory to execute a vision-based parking space detection method disclosed in the first aspect of the embodiments of the present invention.

A fourth aspect of an embodiment of the present invention discloses a computer-readable storage medium that stores a computer program, wherein the computer program causes the computer to execute a vision-based parking space detection method disclosed in the first aspect of the embodiment of the present invention.

A fifth aspect of the embodiments of the present invention discloses a computer program product. When the computer program product runs on a computer, the computer program product is caused to perform part or all of the steps of the method of the first aspect.

A sixth aspect of the embodiments of the present invention discloses an application publishing platform for publishing a computer program product, wherein when the computer program product runs on a computer, the computer is allowed to execute any of the first aspect Part or all steps of a method.

Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

In the embodiment of the present invention, the environmental images on both sides of the captured vehicle are pre-processed to obtain a corresponding bird's-eye view, the parking space frame and the parking space corner point in the bird's-eye view are detected, and one is selected from the parking space frame in the bird's-eye view The target parking space frame, according to the target parking space frame, the target parking space corner point on the target parking space frame is obtained from the parking space corner point, and the real position coordinates corresponding to the target parking space corner point are determined according to the position coordinates of the target parking space corner point in the bird's eye view , According to the real position coordinates corresponding to the corners of the target parking space, determine the target parking space corresponding to the target parking space frame. It can be seen that the implementation of the embodiments of the present invention can improve the accuracy of parking space detection, thereby improving the recognition rate of parking spaces.

BRIEF DESCRIPTION

In order to more clearly explain the technical solutions in the embodiments of the present invention, the drawings required in the embodiments will be briefly described below. Obviously, the drawings in the following description are only some embodiments of the present invention. Those of ordinary skill in the art can obtain other drawings based on these drawings without creative work.

1 is a schematic flowchart of a vision-based parking space detection method disclosed in an embodiment of the present invention;

2 is a schematic flowchart of another vision-based parking space detection method disclosed in an embodiment of the present invention;

3 is a schematic flowchart of another vision-based parking space detection method disclosed in an embodiment of the present invention;

4 is a schematic structural diagram of a parking space detection device disclosed in an embodiment of the present invention;

5 is a schematic structural diagram of another parking space detection device disclosed in an embodiment of the present invention;

6 is a schematic structural diagram of yet another parking space detection device disclosed in an embodiment of the present invention;

7 is a schematic structural diagram of yet another parking space detection device disclosed in an embodiment of the present invention;

FIG. 8 is an application example diagram of a parking space detection process disclosed in an embodiment of the present invention.

detailed description

The technical solutions in the embodiments of the present invention will be described clearly and completely in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present invention.

It should be noted that the terms “first”, “second”, “third”, and “fourth” in the description and claims of the present invention are used to distinguish different objects, not to describe specific order. The terms "including" and "having" and any variations thereof in the embodiments of the present invention are intended to cover non-exclusive inclusions, for example, processes, methods, systems, products, or devices that include a series of steps or units are not necessarily limited to clearly Those steps or units are listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or equipment.

The embodiment of the invention discloses a vision-based parking space detection method and device, which can improve the accuracy of the parking space detection, thereby improving the recognition rate of the parking space. The following is a detailed description with reference to the drawings.

Example one

Please refer to FIG. 1, which is a schematic flowchart of a vision-based parking space detection method disclosed in an embodiment of the present invention. As shown in FIG. 1, the method may include the following steps.

101. The parking space detection device preprocesses the captured environmental images on both sides of the vehicle to obtain a corresponding bird's-eye view.

In the embodiment of the present invention, the parking space detection device photographs the environment on both sides of the vehicle through the B-pillar cameras installed on both sides of the vehicle to obtain the environmental images on both sides of the vehicle, please refer to FIG. 8 together, as shown in FIG. 8 FIG. 8 is an application example diagram of a parking space detection process disclosed in an embodiment of the present invention. When a vehicle performs a parking space detection, a camera device disposed on a B-pillar on both sides of the vehicle is turned on. The camera device detects the environment on both sides of the vehicle. Shoot to obtain environmental images on both sides of the vehicle.

102. The parking space detection device detects the parking space frame and the parking space corner point in the bird's-eye view.

In the embodiment of the present invention, the parking space detection device may use a deep learning algorithm to detect the parking space frame and the parking space corner point in the bird's-eye view, and the embodiment of the present invention is not limited.

As an optional embodiment, the parking space detection device can collect a large number of parking space graphic samples, train to obtain a parking space model based on the parking space graphic samples, and then import the above bird's-eye view into the parking space model to perform convolution operations to generate multiple features. Set the features to detect the parking space frame and parking space corner points from various features. By implementing the above embodiment, the deep learning algorithm is used to detect the parking space frame and the parking space corner point in the bird's-eye view, which can improve the detection accuracy.

Further, as an optional embodiment, the parking space detection device may classify the features, for example, classify the parking space corner points into 0 categories, and classify the parking space frames into 1 category, and then import the above bird's-eye view into the parking space model for volume Product operation is used to generate various features, and the category 0 is extracted as a parking corner and the category 1 is extracted as a parking frame. By implementing the above embodiment, the parking space corner points are divided into 0 categories and the parking space frame is divided into 1 categories, which can increase the recognition speed of the computer and thereby improve the detection efficiency of the parking spaces.

103. The parking space detection device selects a target parking space frame from the parking space frame in the bird's-eye view.

As an optional implementation manner, in step 103, the parking space detection device filters a target parking space frame from the parking space frame in the bird's-eye view, including:

The parking space detection device sorts the width of each parking space frame in the bird's-eye view and counts the ranking of the width of each parking space frame;

The parking space detection device uses the first ranked parking space frame as the target parking space frame.

By implementing the above embodiment, the parking space frame with the largest width is selected as the target parking space frame, and the safety of parking can be improved.

104. The parking space detection device obtains the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame.

In the embodiment of the present invention, the parking space detection device may use a conventional image processing method to filter a target parking space frame from the parking space frame in the bird's-eye view, and remove other parking space corner points except for the target parking space corner point on the target parking space frame.

In the embodiment of the present invention, the number of target parking space corner points on the target parking space frame is at least two, which is not limited in the embodiment of the present invention.

105. The parking space detection device determines the real position coordinates corresponding to the target parking space corner point according to the location coordinates of the target parking space corner point in the bird's-eye view.

In the embodiment of the present invention, formula (1) expresses the mapping relationship between the image coordinate system of the bird's-eye view and the world coordinate system, and the parking space detection device may, according to formula (1), convert the position coordinates of the target parking space corner point in the bird's-eye view Convert to the position coordinates of the target parking space corner point in the world coordinate system, and take the position coordinates of the target parking space corner point in the world coordinate system as the real position coordinates corresponding to the target parking space corner point, formula (1) is as follows:

Where (X _w , Y _w , Z _w ) is the coordinate of the world coordinate system, (x, y) is the coordinate of the image coordinate system of the bird's eye view, Z _c is the height of the camera relative to the ground, and f is the camera's height Focal length, R is the rotation matrix of the camera.

106. The parking space detection device determines the target parking space corresponding to the target parking space frame according to the real position coordinates corresponding to the target parking space corner point.

In the embodiment of the present invention, the parking space detection device determines the center position of the area formed by the target parking space corner point according to the real position coordinates corresponding to the target parking space corner point, and takes the center position of the area formed by the target parking space corner point as the center of the target parking space Location, and place the corner of the target parking space at the corner of the target parking space.

It can be seen that by implementing the method described in FIG. 1, the environmental images on both sides of the vehicle are obtained by shooting, and then the environmental images are converted into a bird's-eye view, and a target parking space frame and the target parking space frame are selected from the parking space frame in the bird's-eye view. For the target parking space corner point, calculating the real position coordinates corresponding to the target parking space corner point and determining the target parking space corresponding to the target parking space frame can improve the accuracy of the parking space detection, thereby improving the recognition rate of the parking space.

Example 2

Please refer to FIG. 2, which is a schematic flowchart of another vision-based parking space detection method disclosed in an embodiment of the present invention. As shown in FIG. 2, the method may include the following steps.

201. The parking space detection device acquires internal parameters and external parameters of the camera device of the vehicle.

In the embodiment of the present invention, the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device, and the external parameters include at least the height of the camera device relative to the ground and the rotation matrix of the camera device, which is not limited in the embodiment of the present invention.

202. The parking space detection device performs distortion correction on the captured environmental images on both sides of the vehicle according to the internal parameters and external parameters of the camera's camera device to obtain corresponding corrected images.

In the embodiment of the present invention, because the image distortion affects the detection of the parking space frame and the parking space corner point, thereby reducing the detection accuracy of the parking space, the image distortion correction needs to be performed before detecting the parking space frame and the parking space corner point. In the embodiment of the present invention, the parking space detection device can correct the distortion of the captured environmental images on both sides of the vehicle according to the internal and external parameters of the camera's camera device, which can be Zhang Zhengyou's checkerboard calibration method, an embodiment of the present invention Not limited.

As an optional embodiment, in step 202, the parking space detection device performs distortion correction on the captured environmental images on both sides of the vehicle according to the internal and external parameters of the camera's camera device to obtain corresponding corrected images, including:

The parking space detection device determines the first correspondence between the environmental images on both sides of the vehicle and the corresponding corrected images based on the internal and external parameters;

The parking space detection device saves the first correspondence to the first look-up table;

The parking space detection device performs distortion correction on the captured environmental images on both sides of the vehicle according to the first lookup table to obtain the corresponding corrected image.

In the embodiment of the present invention, the first correspondence includes the correspondence between the pixels of the environment images captured on both sides of the vehicle and the pixels of the corrected image.

In the embodiment of the present invention, the image captured by the camera device will have a certain degree of distortion, and the corresponding relationship to correct this degree of distortion can be calculated according to the internal parameters and external parameters of the camera device, that is, the parking space detection device acquires the camera device Focal length f, pixel size (d _x ×d _y ) and distortion parameter k=[k ₁ , k ₂ ], and the height Z _{c of the} camera relative to the ground and the rotation matrix R of the camera, and then calculate the corresponding Relationship, where the above corresponding relationship can be expressed by formula (2):

Among them, a coordinate in the world coordinate system of the environment on both sides of the vehicle is (X _w , Y _w , Z _w ), and the coordinate in the image coordinate system of the environment on both sides of the vehicle is (x, y), in the vehicle The coordinates in the pixel coordinate system of the environmental images on both sides are (u, v), the coordinates in the pixel coordinate system of the corrected image are (u′, v′), and the point (u ₀ , v ₀ ) is the light of the camera device The intersection of the axis and the image plane, ie the principal point, Z _c is the height of the camera relative to the ground, d _x is the physical size of the pixel in the x-axis direction, d _y is the physical size of the pixel in the y-axis direction, k ₁ , K ₂ are the first two orders of the distortion parameter k, f is the focal length of the camera, and R is the rotation matrix of the camera.

In the embodiment of the present invention, the parking space detection device may convert the coordinate points (X _w , Y _w , Z _w ) in the environment on both sides of the vehicle in the world coordinate system into the vehicle coordinates in the pixel coordinate system according to formula (1) The pixel coordinate points (u, v) of the environmental image on the side, and then convert each pixel coordinate point (u, v) in the image into the corresponding pixel point (u′, v′) in the corrected image, and Each pixel point (u, v) in the environmental image on the side and the corresponding pixel point (u′, v′) in the corrected image are stored in the first lookup table. Table can find the corresponding pixel point (u′, v′) of each pixel (u, v) in the environmental image on both sides of the vehicle, and then perform distortion correction on the environmental image on both sides of the vehicle, thus Obtain a corrected image.

203. The parking space detection device performs perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view.

In the embodiment of the present invention, the bird's-eye view is a view from the top to the bottom of the actual environment corresponding to the image, because the bird's-eye view view is more conducive to the positioning of the parking space frame and the parking space corner by the parking space detection device. Before the corner of the parking space, the corrected image needs to be converted into a bird's eye view.

As an optional embodiment, in step 203, the parking space detection device performs perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view, including:

The parking space detection device determines the second correspondence between the corrected image and the corresponding bird's-eye view according to the pixels of the corrected image;

The parking space detection device saves the second correspondence to the second look-up table;

The parking space detection device performs perspective transformation on the corrected image according to the second lookup table to obtain a corresponding bird's-eye view.

In the embodiment of the present invention, the second correspondence includes the correspondence between the pixels of the corrected image and the pixels of the bird's-eye view.

In the embodiment of the present invention, the parking space detection device may map the two-dimensional coordinates (u′, v′) of the corrected image in the pixel coordinate system to the three-dimensional coordinates (X _w , Y _w ) in the world coordinate system according to formula (3) Z _w ), divided by the value of Z _w in three-dimensional coordinates to map back to the two-dimensional space between to get the coordinates (u″, v″), so as to obtain a bird’s-eye view, formula (3) is as follows:

Among them, (u′, v′) is the coordinate of the pixel coordinate system of the corrected image, (X _w , Y _w , Z _w ) is the three-dimensional coordinate of the point (u′, v′) in the world coordinate system, (u″ , V″) are the coordinates of the pixel coordinate system of the bird’s-eye view,

Is the transformation matrix, where,

It is the linear transformation of the image, [a ₁₃ a ₂₃ ] ^T is the perspective transformation of the image, and [a ₃₁ a ₃₂ ] is the image translation. In the embodiment of the present invention, given the transformation matrix, the perspective transformation of the corrected image can be performed to obtain a bird's eye view view.

In the embodiment of the present invention, the parking space detection device may convert the pixel coordinate points (u′, v′) in the corrected image in the pixel coordinate system to the pixel coordinate points of the bird's-eye view in the pixel coordinate system according to formula (2) ( u″, v″), and save each pixel point (u′, v′) in the corrected image and the corresponding pixel point (u″, v″) in the bird's-eye view in the second lookup table. When performing perspective transformation on the corrected image, you can directly find the corresponding pixel (u″, v″) of each pixel (u′, v′) in the bird’s-eye view of the corrected image according to the second look-up table, and then correct the The image undergoes perspective transformation to obtain a bird's eye view.

204. The parking space detection device detects the parking space frame and the parking space corner point in the bird's-eye view.

205. The parking space detection device selects a target parking space frame from the parking space frame in the bird's-eye view.

206. The parking space detection device obtains the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame.

207. The parking space detection device detects whether there is an obstacle in the area framed by the target parking space frame; if yes, step 208 is performed; if not, step 209 is performed.

As an optional embodiment, the parking space detection device may select multiple areas at the edge of the target parking space frame and detect the edge density of the multiple areas, and use the edge density of the area corresponding to the maximum edge density as the target edge density, and then determine Whether the target edge density is greater than the specified threshold, if yes, it indicates that there is an obstacle in the area, the parking space detection device performs step 208 to mark the area in the target parking space frame as a non-parking space; if not, the parking space detection device performs step 209 according to the bird's eye view The position coordinates of the target parking space corner point are determined to determine the true position coordinates corresponding to the target parking space corner point. By implementing the above embodiment, it can be determined whether there is an obstacle in the area framed by the target parking space frame selected by the parking space detection device, to avoid collision with the obstacle during the parking process, and the parking safety can be improved.

208. The parking space detection device marks the area within the target parking space frame as a non-parking space.

In the embodiment of the present invention, if there is an obstacle in the area framed by the target parking space frame, it indicates that the parking space corresponding to the target parking space frame cannot accommodate the parking of the vehicle, and the parking space detection device marks the area framed by the target parking space as a non-parking space.

209. The parking space detection device determines the real position coordinates corresponding to the target parking space corner point according to the location coordinates of the target parking space corner point in the bird's-eye view.

210. The parking space detection device determines the target parking space corresponding to the target parking space frame according to the real position coordinates corresponding to the corner of the target parking space.

It can be seen that, compared with the method described in FIG. 1, after implementing the method described in FIG. 2, after determining the target parking space corner point, the parking space detection device detects whether there is an obstacle in the area within the target parking space frame, and if so, it indicates that The parking space cannot be stopped, and the parking space detection device marks the area enclosed by the target parking space frame as a non-parking space, which can prevent the vehicle from hitting obstacles during parking and ensure parking safety. In addition, the method described in FIG. 2 is implemented, and the first and second look-up tables carrying distortion parameters are used to perform distortion correction on the captured environmental images on both sides of the vehicle to obtain corrected images, and then the second look-up tables carrying perspective transformation coefficients are used to correct the images Performing perspective transformation to obtain a bird's-eye view can increase the generation speed of bird's-eye view and improve the detection efficiency of parking spaces.

Example Three

Please refer to FIG. 3, which is a schematic flowchart of another vision-based parking space detection method disclosed in an embodiment of the present invention. As shown in FIG. 3, the method may include the following steps.

301. The parking space detection device acquires the internal parameters and external parameters of the camera device of the vehicle.

302. The parking space detection device determines the third correspondence between the environmental images on both sides of the vehicle and the corresponding bird's-eye view based on the internal parameters and the external parameters.

In the embodiment of the present invention, the third correspondence includes the correspondence between the pixels of the environment images captured on both sides of the vehicle and the pixels of the bird's-eye view.

303. The parking space detection device saves the third correspondence to the third lookup table.

304. The parking space detection device performs bird's eye view transformation on the captured environment images on both sides of the vehicle according to the third lookup table to obtain a corresponding bird's eye view.

In the embodiment of the present invention, the parking space detection device acquires the focal length f, pixel size (d _x ×d _y ) and distortion parameter k = [k ₁ , k ₂ ] of the camera device, and the height of the camera device relative to the ground Z _c and The rotation matrix R of the camera device can then be combined with the above formula (2) and formula (3) to convert the pixel coordinate points (u, v) of the environmental images on both sides of the vehicle in the pixel coordinate system into a bird's eye view in the pixel coordinate system The pixel coordinate points of the view (u″, v″), and save each pixel point (u, v) in the environmental images on both sides of the vehicle and the corresponding pixel point (u″, v″) in the bird’s eye view in the first In the three lookup tables, when it is necessary to perform a bird's eye view transformation on the environmental images on both sides of the vehicle, each pixel point (u, v) in the environmental images on both sides of the vehicle in the bird's eye view can be found directly according to the third lookup table Corresponding to pixel points (u″, v″), and then performing a bird's eye view transformation on the environment images on both sides of the vehicle, thereby obtaining a bird's eye view.

305. The parking space detection device detects the parking space frame and the parking space corner point in the bird's-eye view.

306. The parking space detection device selects a target parking space frame from the parking space frame in the bird's-eye view.

307. The parking space detection device obtains the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame.

308. The parking space detection device detects whether there is an obstacle in the area framed by the target parking space frame; if yes, step 309 is performed; if not, step 310 is performed.

309. The parking space detection device marks the area within the target parking space frame as a non-parking space.

310. The parking space detection device determines the real position coordinates corresponding to the target parking space corner point according to the location coordinates of the target parking space corner point in the bird's-eye view.

311. The parking space detection device determines the target parking space corresponding to the target parking space frame according to the real position coordinates corresponding to the corner of the target parking space.

It can be seen that, compared with the method described in FIG. 1, after implementing the method described in FIG. 3, after determining the target parking space corner point, the parking space detection device detects whether there is an obstacle in the area within the target parking space frame, and if so, it indicates that The parking space cannot be stopped. The parking space detection device marks the area enclosed by the target parking space frame as a non-parking space. In addition, by implementing the method described in FIG. 3, a bird's-eye view is obtained by performing a bird's-eye view transformation on the captured environmental images on both sides of the vehicle through a third look-up table carrying a bird's-eye view transformation coefficient, which can increase the generation speed of the bird's-eye view and improve the efficiency of parking space detection.

Example 4

Please refer to FIG. 4, which is a schematic structural diagram of a parking space detection device disclosed in an embodiment of the present invention. As shown in FIG. 4, the parking space detection device may include:

The preprocessing unit 401 is used to preprocess the captured environment images on both sides of the vehicle to obtain a corresponding bird's-eye view;

The first detection unit 402 is used to detect the parking space frame and the parking space corner point in the bird's-eye view;

The screening unit 403 is used for screening a target parking space frame from the parking space frame in the bird's-eye view;

The obtaining unit 404 is configured to obtain the target parking space corner point on the target parking space frame according to the target parking space frame;

The first determining unit 405 is configured to determine the real position coordinates corresponding to the target parking space corner point according to the position coordinates of the target parking space corner point in the bird's-eye view;

The second determining unit 406 is configured to determine the target parking space corresponding to the target parking space frame according to the real position coordinates corresponding to the target parking space corner point.

In this embodiment of the present invention, the first detection unit 402 may use a deep learning algorithm to detect the parking space frame and the parking space corner point in the bird's-eye view, which is not limited in this embodiment of the present invention.

In the embodiment of the present invention, the screening unit 403 may use a conventional image processing method to filter a target parking space frame from the parking space frame in the bird's-eye view, and exclude other parking space corner points except the target parking space corner point on the target parking space frame.

In the embodiment of the present invention, the number of target parking space corner points on the target parking space frame is at least two, which is not limited in the embodiment of the present invention.

In the embodiment of the present invention, the first determining unit 405 first calculates the position coordinates of the target parking space corner point in the bird's-eye view, and then converts the target parking space angle in the bird's-eye view according to the mapping relationship between the coordinate system of the bird's-eye view and the world coordinate system The position coordinates of the point are converted into the position coordinates of the target parking space corner point in the world coordinate system, and the position coordinates of the target parking space corner point in the world coordinate system are taken as the true position coordinates corresponding to the target parking space corner point.

In the embodiment of the present invention, the second determining unit 406 determines the center position of the area formed by the target parking space corner point according to the real position coordinates corresponding to the target parking space corner point, and takes the center position of the area formed by the target parking space corner point as the target parking space And the corner of the target parking space at the corner of the target parking space.

As an optional implementation manner, the first detection unit 402 can collect a large number of parking space graphic samples, train a parking space model according to the parking space graphic samples, and then import the above bird's-eye view into the parking space model for convolution operation to generate multiple features, A parking space frame and a parking space corner point are detected from various characteristics according to preset characteristics. Implementing the above embodiment, using a deep learning algorithm to detect the parking space frame and the parking space corner point in the bird's-eye view, can improve the detection accuracy.

Further, as an optional implementation manner, the first detection unit 402 may classify features, for example, classify parking space corner points into 0 categories, and classify parking space frames into 1 category, and then import the above bird's-eye view into the parking space model A convolution operation is performed to generate a variety of features, and the category 0 is extracted as a parking corner and the category 1 is extracted as a parking frame. By implementing the above embodiment, the parking space corner points are divided into 0 categories and the parking space frame is divided into 1 categories, which can increase the recognition speed of the computer and thereby improve the detection efficiency of the parking spaces.

As an optional implementation manner, the screening unit 403 screens a target parking space frame from the parking space frame in the bird's-eye view, including:

The filtering unit 403 sorts the width of each parking space frame in the bird's-eye view and counts the ranking of the width of each parking space frame;

The screening unit 403 uses the first ranked parking space frame as the target parking space frame.

By implementing the above embodiment, the parking space frame with the largest width is selected as the target parking space frame, and the safety of parking can be improved.

It can be seen that by implementing the parking space detection device described in FIG. 4, the environmental images on both sides of the vehicle are obtained by shooting, and then the environmental images are converted into a bird's eye view, and a target parking space frame and the target parking space frame are selected from the parking space frame in the bird's eye view On the target parking space corner point, the real position coordinates corresponding to the target parking space corner point are calculated and the target parking space corresponding to the target parking space frame is determined, which can improve the accuracy of the parking space detection and thereby increase the recognition rate of the parking space.

Example 5

Please refer to FIG. 5, which is a schematic structural diagram of another parking space detection device disclosed in an embodiment of the present invention. Among them, the parking space detection device shown in FIG. 5 is further optimized by the parking space detection device shown in FIG. 4. Compared with the parking space detection device shown in FIG. 4, the parking space detection device shown in FIG. 5 may further include:

The pre-processing unit 401 includes:

The distortion correction subunit 4011 is configured to perform distortion correction on the captured environmental images on both sides of the vehicle according to the internal parameters and external parameters of the camera's camera device to obtain corresponding corrected images;

The perspective transformation subunit 4012 is configured to perform perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view.

The first parameter acquisition unit 407 is used to determine before the distortion correction subunit 4011 performs distortion correction on the captured environmental images on both sides of the vehicle according to the internal and external parameters of the camera device of the vehicle to obtain the corresponding corrected image The internal and external parameters of the camera device of the vehicle;

In the embodiment of the present invention, the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device, and the external parameters include at least the height of the camera device relative to the ground and the rotation matrix of the camera device, which is not limited in the embodiment of the present invention.

As an optional embodiment, the distortion correction subunit 4011 performs distortion correction on the captured environmental images on both sides of the vehicle according to the internal and external parameters of the camera's camera device to obtain corresponding corrected images, including:

The first determination module 40111 is configured to determine the first correspondence between the environmental images on both sides of the vehicle and the corresponding corrected images according to the internal parameters and external parameters determined by the first parameter acquisition unit 407;

The first saving module 40112 is used to save the first correspondence to the first lookup table;

The distortion correction module 40113 is configured to perform distortion correction on the captured environment images on both sides of the vehicle according to the first lookup table to obtain a corresponding corrected image.

In the embodiment of the present invention, the first correspondence includes the correspondence between the pixels of the environment images captured on both sides of the vehicle and the pixels of the corrected image.

As an optional embodiment, the perspective transformation subunit 4012 performs perspective transformation on the corrected image according to the pixels of the corrected image to obtain the corresponding bird's-eye view, including:

The second determining module 40121 is configured to determine the second correspondence between the corrected image and the corresponding bird's-eye view according to the pixels of the corrected image;

The second saving module 40122 is used to save the second correspondence to the second lookup table;

The perspective transformation module 40123 is configured to perform perspective transformation on the corrected image according to the second lookup table to obtain a corresponding bird's-eye view.

In the embodiment of the present invention, the second correspondence includes the correspondence between the pixels of the corrected image and the pixels of the bird's-eye view.

The second detection unit 408 is used for the acquisition unit 404 to acquire the target parking space corner on the target parking space frame from the parking space corner according to the target parking space frame, and the first determination unit 405 according to the target parking space corner in the bird's-eye view Before determining the real position coordinates corresponding to the corner of the target parking space, detect whether there is an obstacle in the area framed by the target parking space frame;

The marking unit 409 is used to mark the area in the target parking space frame as a non-parking space when the second detection unit 408 detects that there is an obstacle in the area in the target parking space frame;

The first determining unit 405 is specifically configured to determine the target parking space corner according to the position coordinates of the target parking space corner in the bird's-eye view when the second detection unit 408 detects that there is no obstacle in the area framed by the target parking space frame Corresponding real position coordinates.

As an optional implementation manner, the second detection unit 408 may screen multiple areas at the edge of the target parking space frame and detect the edge density of the multiple areas, and use the edge density of the area corresponding to the maximum edge density as the target edge density, Next, determine whether the target edge density is greater than the specified threshold. If it is, it indicates that there is an obstacle in the area. The marking unit 409 marks the area within the target parking frame as a non-parking space; if not, the first determining unit 405 determines The position coordinates of the target parking space corner point determine the true position coordinates corresponding to the target parking space corner point.

By implementing the above embodiment, it can be determined whether there is an obstacle in the area framed by the target parking space frame selected by the parking space detection device, to avoid collision with the obstacle during the parking process, and the parking safety can be improved.

It can be seen that, compared with the parking space detection device described in FIG. 4, after implementing the parking space detection device described in FIG. 5, after determining the target parking space corner point, it is detected whether there is an obstacle in the area within the target parking space frame, and if so, it indicates that The parking space cannot be stopped. Mark the area enclosed by the target parking space frame as a non-parking space, which can avoid the vehicle from hitting obstacles during parking and ensure parking safety. In addition, the parking space detection device described in FIG. 5 is implemented, and the first and second look-up tables carrying distortion parameters are used to perform distortion correction on the captured environmental images on both sides of the vehicle to obtain corrected images, and then the second look-up table carrying perspective transformation coefficients Correcting the image for perspective transformation to obtain a bird's eye view can increase the speed of bird's eye view generation and improve the efficiency of parking space detection.

Example Six

Please refer to FIG. 6, which is a schematic structural diagram of another parking space detection device disclosed in an embodiment of the present invention. Among them, the parking space detection device shown in FIG. 6 is further optimized by the parking space detection device shown in FIG. 4. Compared with the parking space detection device shown in FIG. 4, the parking space detection device shown in FIG. 6 may further include:

The second parameter obtaining unit 410 is used to obtain the internal parameters and external parameters of the camera device of the vehicle before the preprocessing unit 401 preprocesses the captured environment images on both sides of the vehicle to obtain the corresponding bird's-eye view;

The pre-processing unit 401 includes:

The determination subunit 4013 is configured to determine the third correspondence between the environmental images on both sides of the vehicle and the corresponding bird's-eye view according to the internal and external parameters determined by the second parameter acquisition unit 410;

Saving subunit 4014, used to save the third correspondence to the third lookup table;

The bird's eye view transformation subunit 4015 is configured to perform bird's eye view transformation on the captured environment images on both sides of the vehicle according to the third lookup table to obtain a corresponding bird's eye view.

In the embodiment of the present invention, the third correspondence includes the correspondence between the pixels of the environment images captured on both sides of the vehicle and the pixels of the bird's-eye view.

The second detection unit 408 is used for the acquisition unit 404 to acquire the target parking space corner on the target parking space frame from the parking space corner according to the target parking space frame, and the first determination unit 405 according to the target parking space corner in the bird's-eye view Before determining the real position coordinates corresponding to the corner of the target parking space, detect whether there is an obstacle in the area framed by the target parking space frame;

The marking unit 409 is used to mark the area in the target parking space frame as a non-parking space when the second detection unit 408 detects that there is an obstacle in the area in the target parking space frame;

The first determining unit 405 is specifically configured to determine the target parking space corner according to the position coordinates of the target parking space corner in the bird's-eye view when the second detection unit 408 detects that there is no obstacle in the area framed by the target parking space frame Corresponding real position coordinates.

It can be seen that, compared with the parking space detection device described in FIG. 4, after implementing the parking space detection device described in FIG. 6, after determining the target parking space corner point, it is detected whether there is an obstacle in the area within the target parking space frame, and if so, it indicates that The parking space cannot be stopped. Mark the area enclosed by the target parking space frame as a non-parking space, which can avoid the vehicle from hitting obstacles during parking and ensure parking safety. In addition, by implementing the parking space detection device described in FIG. 6 and performing a bird's eye view transformation on the captured environmental images on both sides of the vehicle by carrying a third look-up table of bird's eye view transformation coefficients to obtain a bird's eye view, the speed of generating bird's eye view can be improved and the parking space detection can be improved effectiveness.

Example 7

Please refer to FIG. 7, which is a schematic structural diagram of another parking space detection device disclosed in an embodiment of the present invention. As shown in FIG. 7, the parking space detection device may include:

A memory 701 storing executable program code;

A processor 702 coupled with the memory 701;

Wherein, the processor 702 calls the executable program code stored in the memory 701 to execute any one of the vision-based parking space detection methods shown in FIGS. 1 to 3.

An embodiment of the present invention discloses a computer-readable storage medium that stores a computer program, where the computer program causes the computer to execute any of the vision-based parking space detection methods of FIGS. 1 to 3.

An embodiment of the present invention also discloses an application publishing platform, wherein the application publishing platform is used to publish a computer program product, wherein, when the computer program product runs on the computer, the computer is caused to perform part of the method as in the above method embodiments Or all steps.

In various embodiments of the present invention, it should be understood that the size of the sequence numbers of the above processes does not mean that the order of execution is necessarily inevitable. The execution order of each process should be determined by its function and inherent logic, and should not be implemented by the present invention. The implementation process of the examples constitutes no limitation.

The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be object units, that is, may be located in one place, or may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objective of the solution of this embodiment. In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or software function unit.

In the embodiment provided by the present invention, it should be understood that "B corresponding to A" indicates that B is associated with A, and B can be determined according to A. However, it should also be understood that determining B based on A does not mean determining B based on A alone, and B may also be determined based on A and/or other information.

Those of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by a program instructing related hardware. The program may be stored in a computer-readable storage medium, and the storage medium includes read-only Memory (Read-Only Memory, ROM), Random Memory (Random Access, Memory, RAM), Programmable Read-only Memory (PROM), Erasable Programmable Read-Only Memory (Erasable Programmable Read Only Only Memory, EPROM), One-time Programmable Read-Only Memory (OTPROM), electronically erasable rewritable read-only memory (Electrically-Erasable Programmable Read-Only Memory, EEPROM), compact disc (Compact Disc) Read-Only Memory (CD-ROM) or other optical disk storage, magnetic disk storage, magnetic tape storage, or any other medium readable by a computer that can be used to carry or store data.

The vision-based parking space detection method and device disclosed in the embodiments of the present invention are described in detail above. Specific examples are used in this article to explain the principles and implementation of the present invention. The descriptions of the above embodiments are only used to help understanding The method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation and scope of application. In summary, the content of this specification should not It is understood as a limitation to the present invention.

Claims (10)

1. A vision-based parking space detection method, which includes:

   Preprocess the environmental images on both sides of the captured vehicle to obtain the corresponding bird's-eye view;

   Detecting a parking space frame and a parking space corner point in the bird's-eye view;

   Selecting a target parking space frame from the parking space frame in the bird's-eye view;

   Obtaining the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame;

   Determine the real position coordinates corresponding to the target parking space corner point according to the position coordinates of the target parking space corner point in the bird's-eye view;

   The target parking space corresponding to the target parking space frame is determined according to the real position coordinates corresponding to the target parking space corner point.
2. The method according to claim 1, wherein the preprocessing the captured environmental images on both sides of the vehicle to obtain a corresponding bird's-eye view includes:

   According to the internal parameters and external parameters of the camera's camera device, perform distortion correction on the captured environmental images on both sides of the vehicle to obtain corresponding corrected images;

   Perform perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view.
3. The method according to claim 2, characterized in that, in the internal parameters and external parameters of the camera device of the vehicle, the captured environmental images on both sides of the vehicle are subjected to distortion correction to obtain a corresponding corrected image Previously, the method also included:

   Obtain the internal and external parameters of the camera's camera device; the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the camera Device rotation matrix;

   According to the internal parameters and external parameters of the camera's camera device, the captured environmental images on both sides of the vehicle are subjected to distortion correction to obtain corresponding corrected images, including:

   According to the internal parameter and the external parameter, determine a first correspondence between the environment images on both sides of the vehicle and corresponding corrected images; the first correspondence includes the captured environment on both sides of the vehicle Correspondence between pixels of the image and pixels of the corrected image;

   Save the first correspondence to the first lookup table;

   According to the first look-up table, perform distortion correction on the captured environmental images on both sides of the vehicle to obtain the corrected image;

   The performing perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view includes:

   According to the pixels of the corrected image, determine a second correspondence between the corrected image and the corresponding bird's-eye view; the second correspondence includes the pixels of the corrected image and the pixels of the bird's-eye view Correspondence between

   Save the second correspondence to the second lookup table;

   According to the second lookup table, perform a perspective transformation on the corrected image to obtain the bird's-eye view.
4. The method according to claim 1, characterized in that before the preprocessing of the captured environment images on both sides of the vehicle to obtain a corresponding bird's-eye view, the method further comprises:

   Obtain the internal and external parameters of the camera's camera device; the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the camera Device rotation matrix;

   And, the preprocessing of the environmental images on both sides of the captured vehicle to obtain a corresponding bird's-eye view includes:

   According to the internal parameter and the external parameter, determine a third correspondence between the environment images on both sides of the vehicle and the corresponding bird's-eye view; the third correspondence includes the captured environment on both sides of the vehicle Correspondence between pixels of the image and pixels of the bird's-eye view;

   Save the third correspondence to the third lookup table;

   According to the third lookup table, perform a bird's eye view transformation on the captured environment images on both sides of the vehicle to obtain the bird's eye view.
5. The method according to any one of claims 1 to 4, characterized in that after the target parking space corner on the target parking space frame is obtained from the parking space corner according to the target parking space frame, and Before determining the real position coordinates corresponding to the target parking space corner point according to the position coordinates of the target parking space corner point in the bird's-eye view, the method further includes:

   Detecting whether there is an obstacle in the area framed by the target parking space frame;

   If there is an obstacle in the area framed by the target parking space frame, mark the area framed by the target parking space frame as a non-parking space;

   If there is no obstacle in the area framed by the target parking space frame, the step of determining the real position coordinates corresponding to the target parking space corner point according to the location coordinates of the target parking space corner point in the bird's-eye view is performed .
6. A parking space detection device, characterized in that it includes:

   The pre-processing unit is used to pre-process the environment images on both sides of the vehicle to obtain the corresponding bird's-eye view;

   A first detection unit, configured to detect a parking space frame and a parking space corner point in the bird's-eye view;

   A screening unit for screening a target parking space frame from the parking space frame in the bird's-eye view;

   An obtaining unit, configured to obtain a target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame;

   A first determining unit, configured to determine the real position coordinates corresponding to the target parking space corner point according to the position coordinates of the target parking space corner point in the bird's-eye view;

   The second determining unit is configured to determine the target parking space corresponding to the target parking space frame according to the real position coordinates corresponding to the target parking space corner point.
7. The parking space detection device according to claim 6, wherein the preprocessing unit includes:

   The distortion correction subunit is used to perform distortion correction on the captured environmental images on both sides of the vehicle according to the internal parameters and external parameters of the camera device of the vehicle to obtain corresponding corrected images;

   The perspective transformation subunit is configured to perform perspective transformation on the corrected image according to the pixels of the corrected image to obtain a corresponding bird's-eye view.
8. The parking space detection device according to claim 7, wherein the parking space detection device further comprises:

   The first parameter acquisition unit is used to perform distortion correction on the captured environmental images on both sides of the vehicle according to the internal and external parameters of the camera's camera device of the distortion correction sub-unit to obtain the corresponding corrected image To obtain the internal and external parameters of the camera's camera device; the internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the The rotation matrix of the camera device;

   The distortion correction subunit includes:

   A first determining module, configured to determine the first correspondence between the environmental images on both sides of the vehicle and the corresponding corrected images based on the internal parameters and the external parameters; the first correspondence includes the captured Correspondence between pixels of the environmental images on both sides of the vehicle and pixels of the corrected image;

   A first saving module, configured to save the first correspondence to the first lookup table;

   A distortion correction module, configured to perform distortion correction on the captured environmental images on both sides of the vehicle according to the first lookup table to obtain the corrected image;

   The perspective transformation subunit includes:

   A second determining module, configured to determine a second correspondence between the corrected image and the corresponding bird's-eye view according to the pixels of the corrected image; the second correspondence includes the pixels of the corrected image and all The correspondence between the pixels of the bird's-eye view;

   A second saving module, configured to save the second correspondence to the second lookup table;

   The perspective transformation module is configured to perform perspective transformation on the corrected image according to the second lookup table to obtain the bird's-eye view.
9. The parking space detection device according to claim 6, wherein the parking space detection device further comprises:

   A second parameter acquisition unit, configured to acquire the internal parameters and external parameters of the camera device of the vehicle before the preprocessing unit preprocesses the captured environment images on both sides of the vehicle to obtain the corresponding bird's-eye view; The internal parameters include at least the focal length, pixel size, and distortion parameters of the camera device; the external parameters include at least the height of the camera device relative to the ground and the rotation matrix of the camera device;

   The pre-processing unit includes:

   A determination subunit, configured to determine a third correspondence between the environmental images on both sides of the vehicle and the corresponding bird's-eye view based on the internal parameter and the external parameter; the third correspondence includes the A correspondence between pixels of environmental images on both sides of the vehicle and pixels of the bird's-eye view;

   A saving subunit, configured to save the third correspondence to a third lookup table;

   The bird's eye view transformation subunit is configured to perform bird's eye view transformation on the captured environment images on both sides of the vehicle according to the third lookup table to obtain the bird's eye view.
10. The parking space detection device according to any one of claims 6 to 9, wherein the parking space detection device further comprises:

    The second detection unit is used for the acquisition unit to acquire the target parking space corner point on the target parking space frame from the parking space corner point according to the target parking space frame, and the first determination unit according to the Detecting the position coordinates of the target parking space corner point in the bird's-eye view, before determining the real position coordinates corresponding to the target parking space corner point, detecting whether there is an obstacle in the area within the target parking space frame;

    A marking unit, configured to mark the area in the target parking space frame as a non-parking space when the second detection unit detects that there is an obstacle in the area in the target parking space frame;

    The first determining unit is specifically configured to determine the position of the corner point of the target parking space in the bird's-eye view when the second detection unit detects that there is no obstacle in the area framed by the target parking space frame Coordinates, determine the real position coordinates corresponding to the target parking space corner point.

Images